1. Field of the Invention
The present invention relates to a manufacturing process change control apparatus and a manufacturing process change control method, and more specifically to a manufacturing process change control apparatus and manufacturing process change control method that are suitable for use in a manufacturing control system for the manufacturing processes of semiconductor devices.
2. Description of Related Art
In the past, control of manufacturing processes in a factory was generally done in real time, using control diagrams and the like, and it was necessary to quickly determine the cause in the case in which an abnormality was discovered in any of the trends collected for the controlled processes.
However, in a factory having many processes and manufacturing processes in which the same equipment is used a large number of times, such as in a semiconductor factory, the causes of disturbance to the manufacturing processes interact in a complex manner, so that there are many cases in which a large amount of time is expended in discovering the basic cause.
In particular with regard to such changed items as equipment maintenance, which is performed in asynchronously with respect to manufacturing, it is extremely difficult to control the influence that each change item has on the manufacturing process.
Because of the above-noted situation, in many manufacturing factories, such as semiconductor factories, there are efforts being made to automate the manufacturing processes and the system of managing the manufacturing processes, for the purpose of not only improving quality and reliability, but also of increasing the usage efficiency of manufacturing equipment and lowering the manufacturing cost.
For example, in the case of controlling the quality of a product, it is necessary to control all of the manufacturing processes.
In particular with regard to a semiconductor manufacturing factory, because of the large number of and complexity of the manufacturing processes, the observation of the quality of each and every process requires a large amount of time, this imposing limits to control performed by humans.
Additionally, because of the complex interrelationships between processes, it is difficult to discover the cause of a disturbance or abnormality in manufacturing data by merely managing each individual process, this approach necessarily becoming an insufficient one.
Because of this situation, there has been an attempt to configure a manufacturing factory to enable automated control and improvement of the overall manufacturing process, one example with regard to process control system aimed at automation of quality control in a semiconductor manufacturing factory being a manufacturing process quality problem handling system that is disclosed, for example, in the Japanese Unexamined Patent Publication (KOKAI) No. 8-202775.
In this system, each of the manufacturing data that are collected for one and the same control unit by a process data collection apparatus is compared by means of a collected data checking apparatus with respect to control values and control upper limits and control lower limits which are established in accordance with the product specifications, a comparison being made to determine whether or not values are within a set control range, and if this range is exceeded, an alarm is issued, thereby notifying a worker of the abnormality.
Simultaneously with the above, sampling is used to calculate average values and dispersions, these trends in collected data being checked, this actual data being used as the basis for automatic resetting of more severe control ranges, thereby performing strict quality control so as to maintain product quality at a high level, this being proposed as one method of performing strict control of a general process control system and quality.
In the above-described manufacturing processing quality abnormality handling system, the automatic specification judgment for each process is basically only making a judgment as the good or bad quality of the results from the immediately previous manufacturing process, so that even if an alarm is issued, this is only notification that some type of abnormality has occurred in a manufacturing process, that is, that a process exceeds a control limit, and this does not provide notification of the location in the manufacturing process or the cause of the abnormality.
Therefore, although it is possible to use this alarm to stop manufacturing and minimize the abnormal products that occur, time is required to actually discover the cause of the abnormality, during which, because the manufacturing process is stopped, a great loss is incurred in manufacturing.
The reason the above-noted problem occurs is because the information with regard to the cause of the abnormality is ignored.
In general, causes of such abnormalities are various, and can be envisioned as including suddenly occurring equipment problems, human errors such as improper settings of conditions and improper maintenance, and variations between lots of raw materials or changes in quality of purchased materials.
Of the above, while it is impossible to avoid suddenly occurring equipment problems with a conventional control system, there is generally some information with regard to the other problems, and the reason that manufacturing process quality problem handling systems of the past could not teach of the cause of an abnormality was that absolutely no consideration was given to the relationship to this information.
The present invention was made in view of the above-described situation, and has an object the provision of a manufacturing process change control apparatus and a manufacturing process change control method which, when a change point occurs in a manufacturing process, performs a statistical comparison judgment between the information in the manufacturing process for each changed item after the change point with the same information before the change point, thereby providing a direct and quick grasp of the cause of the abnormality in manufacturing data and enabling quick corrective action.